The New Puppeteers of NHS Care: How US Tech Is Moving Inside Britain’s Health Service

Walk through the technology behind the modern NHS and an uncomfortable pattern emerges. Epic and Oracle sit inside major hospital record systems. Palantir leads the consortium behind NHS England’s Federated Data Platform. Amazon Web Services supports healthcare cloud infrastructure. Dell supplies enterprise technology across the public sector. Microsoft is moving even further into everyday NHS operations, with NHS England rolling out Microsoft 365 Copilot to 505,000 clinicians and support staff during 2026. In 2023, UnitedHealth Group, parent of Optum, also completed its £1.2 billion acquisition of EMIS, one of Britain’s most important primary care technology companies. These companies perform very different roles, but the direction is striking: increasingly important layers of British healthcare are becoming intertwined with large US technology corporations.

The irony is that the NHS is nothing like the healthcare economy in which many of these companies grew up. Britain’s model is much closer to Spain: virtually universal coverage, largely funded through taxation and predominantly delivered through the public sector, free at the point of care. The American system developed through a far more fragmented mixture of private insurance, public programmes, reimbursement mechanisms and historically powerful fee for service incentives. Software does not arrive carrying an American flag, and good technology can cross borders brilliantly. But technology is rarely neutral. Platforms are shaped by the markets, workflows and commercial incentives around which they were built. Britain should therefore be asking a question it has largely avoided: why is a European style public health service becoming so dependent on an American technology ecosystem?

This did not happen overnight. The foundations were laid in 2002 with the National Programme for IT, the extraordinary £12.4 billion attempt to digitise the NHS through national infrastructure and large external suppliers. The programme struggled and was eventually dismantled, but something survived: the instinct that enormous NHS technology problems require enormous technology companies. Then came cloud first government policy, the acceleration of digital services during COVID, the expansion of electronic patient records and, finally, the AI era. By 2023, NHS England had awarded a Palantir led consortium the Federated Data Platform contract, with investment of up to £330 million over seven years. Epic is now used by major NHS organisations, while Oracle Millennium operates across large hospital groups.

None of this means American companies “own” NHS patient data. They do not. NHS organisations remain responsible for the information they control, and national programmes carry contractual and governance protections around its use. The more troubling issue is dependency. A company does not need to own the hospital or legally own its records to become extraordinarily important to how that hospital functions. If the clinical record, cloud infrastructure, analytics environment, productivity software and increasingly the AI assistant used by hundreds of thousands of NHS workers depend on a small group of global suppliers, switching becomes harder. Integration becomes power. Architecture becomes power. The cost of leaving becomes power.

Microsoft makes that question particularly immediate. The NHS is not merely buying desktop software. Its 2026 agreement is intended to place Microsoft 365 Copilot in the hands of more than half a million NHS workers, with access to tools capable of helping draft patient letters, analyse information, support discharge processes, build rotas, assist with bed management and produce organisational analysis. NHS England says the technology could release substantial clinical and administrative time. That could be transformative. But it also means another layer of everyday NHS work, from clinical administration to management decision making, will increasingly pass through a US technology platform.

Then comes the wider AI question. The NHS Single Patient Record is intended to bring information together and eventually make it accessible through the NHS App. Add predictive algorithms, remote monitoring, automated scheduling, workforce analytics and AI agents, and the NHS could become observable in extraordinary detail. That can improve care dramatically. It can also create something unprecedented: infrastructure capable of seeing where patients go, what clinicians do, how hospitals perform, where capacity sits and how decisions are being made across increasingly connected services.

That is not an accusation of improper surveillance. It is a governance problem waiting to be understood.

The same digital architecture capable of creating a safer, smarter NHS is also capable of creating a far more measurable NHS. Britain therefore needs to decide who can see what, who can process what, how long data is retained, which systems can connect and what happens when the NHS wants to move supplier. Those decisions should not be made accidentally through procurement.

There is an economic question too. Britain is investing billions in digitising one of the world’s largest health systems. Every major procurement decision helps determine where the resulting capability, intellectual property, engineering expertise and enterprise value accumulate. The alternative is not technological nationalism or buying an inferior British product because it carries a Union Jack. It is creating procurement conditions in which British and European companies can actually compete: open standards, smaller modular contracts, genuine interoperability, mandatory portability and architectures that do not require a new entrant to fight through an incumbent platform simply to reach NHS data.

The NHS therefore faces a choice far bigger than Epic versus Oracle, Microsoft versus another cloud provider or Palantir versus somebody else. It must decide whether digital transformation creates sovereign capability or simply a more sophisticated form of dependency. American technology can and should have a place in British healthcare when it is the best technology available. But no supplier should become indispensable, no patient record should become practically immovable and no NHS organisation should discover that changing direction requires permission from the architecture it purchased years earlier.

The new puppeteers of healthcare may not control the NHS today. The danger is building a system in which, one contract, one platform and one AI deployment at a time, Britain eventually discovers that it cannot move without them.